Not a great article but the figures are what I find interesting. It seems these 'low power' distributions are often run on guesswork, prejudice and rules of thumb. There's no scientific framework for measuring the actual impact of certain decisions.
If such a framework existed then any Linux distro could use it and it would catch stupid bugs in shared infrastructure that suddenly starts eating memory or CPU and let people focus on areas ripe for improvement. Similar things have happened with identifying power hogs, browser memory usage, or things slowing down the boot process. Often there are many stupid bugs that can easily be fixed without having to throw everything out and start again.
LXDE based Lubuntu seems to do better in this regard than the XFCE based Xubuntu but I'd still like to see it formalised to the degree that you could make informed decisions about the tradeoff between different browsers or whatever and see standard Ubuntu get 10% "faster" or "lighter" with every release rather than slowly bloat up till something breaks and people start to notice. The less L/Xubuntu has to diverge from the mainstream the better for everyone.
Terrible article. Whines about choice, then concludes the latest choice is compelling. Whines about fragmentation in Ubuntu, then at the very end shows that he does understand that it is in fact all Ubuntu and all these "variants" are just different installed packages. Spends about half of an article with some interesting tidbits whining in general.
And in the end, his conclusion is wrong, too. The different between a 200MB and 300MB memory load is not all that significant; if your machine can run, it can run the other, with a bit of swap. (In either case you're looking at swap anyhow, and bear in mind that I'm accounting for the fact we're talking about soft limits, not hard ones. 300MB actual MB is different than 200 actual MB on a 256MB system, but memory doesn't work that way, it is always fluctuating up and down. 200-ish isn't all that different than 300-ish.) If you've got 256MB of RAM, you'll pretty much be in the same dire straights with either system, and comfy at 512MB; if you only have 64 or 128MB, you're screwed with both choices. So, this is a pretty narrow window of applicable results.
It should be broken into two articles, one about memory consumption, and one which contains all the whining. Then I could ignore the second one.
What about Fluxbuntu? I thought that distro's goal was to be lightweight and put little demand on system resources? In the article he mentions the distro, but instead draws a comparison between Ubuntu/Xubuntu/Lubuntu. What?
5 comments
[ 3.5 ms ] story [ 28.9 ms ] threadIf such a framework existed then any Linux distro could use it and it would catch stupid bugs in shared infrastructure that suddenly starts eating memory or CPU and let people focus on areas ripe for improvement. Similar things have happened with identifying power hogs, browser memory usage, or things slowing down the boot process. Often there are many stupid bugs that can easily be fixed without having to throw everything out and start again.
LXDE based Lubuntu seems to do better in this regard than the XFCE based Xubuntu but I'd still like to see it formalised to the degree that you could make informed decisions about the tradeoff between different browsers or whatever and see standard Ubuntu get 10% "faster" or "lighter" with every release rather than slowly bloat up till something breaks and people start to notice. The less L/Xubuntu has to diverge from the mainstream the better for everyone.
And in the end, his conclusion is wrong, too. The different between a 200MB and 300MB memory load is not all that significant; if your machine can run, it can run the other, with a bit of swap. (In either case you're looking at swap anyhow, and bear in mind that I'm accounting for the fact we're talking about soft limits, not hard ones. 300MB actual MB is different than 200 actual MB on a 256MB system, but memory doesn't work that way, it is always fluctuating up and down. 200-ish isn't all that different than 300-ish.) If you've got 256MB of RAM, you'll pretty much be in the same dire straights with either system, and comfy at 512MB; if you only have 64 or 128MB, you're screwed with both choices. So, this is a pretty narrow window of applicable results.
It should be broken into two articles, one about memory consumption, and one which contains all the whining. Then I could ignore the second one.